Iranian Journal of Materials Science & Engineering Vol. 17, No. 3, September 2020

1

1 . INTRODUCTION

Glass foam has a unique combination of prop-erties such as low density, suitable rigidity, ther-mal insulation, rodent, insect, water and steam resistance. Commercial glass foams exhibit po-rosity, apparent density, and compressive strength values of approximately 85 – 95 vol%, 0.1 – 0.3 g/cm3, and 0.4 – 6 MPa, respectively[1].

Manufacturing glass foam via recycling sili-cate industrial wastes has been popular in the last two decades. However, the reported mechanical strengths of these materials have not been satis-factorily [2-15]. To improve such a deficiency and to achieve more control on foaming behavior, re-searches have been conducted on manufacturing glass-ceramic foams[5,8,11,16]. But, almost none of them, namely cathode ray tube (CRT) and so-da-lime glasses have met the required mechani-cal properties with proper relative density. This is mainly because the mentioned systems are hardly inclined to crystallize [5, 8, 11].

Transition metal oxides containing glasses seem to be appropriate candidates to make up the above-mentioned deficiency. For instance, it is said that iron oxide affects both nucleation and crystallization of silicate glasses [17-19]. Based

High Strength Pyroxene-Based Glass- Ceramic Foams in the Presence of Fe2O3

P. Shahsavari, B. Eftekhari Yekta* and V. K. Marghussian * beftekhari@iust.ac.ir

Received: April 2020 Revised:July 2020 Accepted: August 2020

* School of Metallurgy and Materials Engineering, Department of Ceramics, Iran University of Science and Technol-ogy, Tehran, Iran. DOI: 10.22068/ijmse.17.3.1

Abstract: Strong glass-ceramic foams with a compressive strength of 20 MPa were prepared by adding various amounts of Fe2O3 to a soda lime-based glass composition, using SiC as a foaming agent. The foams were prepared by heat treatment of the compacted samples in the range of 750–950°C and various soaking times. The crystallization behavior of the samples was investigated by Simultaneous Thermal Analysis (STA), Scanning Electron Microscope, and X-Ray Diffractometer (XRD). Based on the results, solid solutions of pyroxene groups were crystallized by the surface mechanism, between 730 and 900˚C, and their amounts increased with the increase in the added iron oxide. In addition, we found that Fe2O3 neither acts as a nucleant for pyroxene nor as an oxidizer for SiC. The results also showed that the compressive strength as well as the crystallization behavior of the foams were influenced by the presence of the SiC particles.

Keywords: Glass-Ceramic, Foam, Pyroxene.

on various reports, multiple ion valences of iron along with alkaline oxides facilitate pyroxene formation and improve the abrasion resistance, mechanical strength and chemical durability of glasses [18, 20, 21].

Since the foaming performance of soda-lime glass in the presence of SiC particles has been approved previously, various glasses containing similar ratios of a soda-lime glass constituent, and amounts of iron oxide were melted, mixed with SiC particles, and their crystallization behavior, density and strength were evaluated. The cho-sen compositions were simple simulations of the well-known Iranian copper slag by-product. The authors believed that the results of this experiment could help us to have a better evaluation of the glass foams, which will produce from copper slag.

2 . EXPERIMENTAL PROCEDURE

The chemical compositions of the glasses that will be denoted as Gx in this article are shown in Table1. Here “X” indicates the weight of the part of iron oxide, which was present in each glass composition. In fact, the glasses are formulated based on a soda-lime glass composition; but, con-tain various amounts of Fe2O3.

RESEARCH PAPER

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

2

P. Shahsavari, et al.

The glass batches were melted in zircon cruci-ble in an electric furnace at 1450 °C for 1 h. Then, the melts were quenched into cooled water. The frits were milled in a planetary mill for 20 min. The mean particle size of the frits was approxi-mately 20 µm. The ground powders were mixed thoroughly with SiC powder (reagent grade, Sig-ma–Aldrich, and particle size <45 µm) and suit-able amounts of PVA solution.

The amounts of SiC, which were added to the aforementioned ground frits, were 1, 2, 3, and 4%wt.

The mixed batches were uniaxially pressed un-der 80 MPa pressure using a hydraulic press into a 20 mm diameter cylindrical steel die. The samples were then heat-treated between 700 and 950 °C with a soaking time of 25 min and cooled to room temperature in the furnace. The heating rate was 10 °C/min.

The crushing strength was determined using specimens with 5mm×10mm×10mm dimensions.

An Instron Universal Testing Machine with a crosshead speed of 1 mm/min was used in these experiments. At least three specimens were evalu-ated for each composition.

The bulk density of the cubic samples was ob-tained by measuring their dimensions and their true density was determined via the pycnometer. The thermal analysis was carried out by simulta-neous thermal analysis (NETZSCH 409) in the air using a heating rate of 10°C/min.

The crystallinity of the samples was investigated by an X-ray diffractometer (Philips PW1800). The micro-structure of the samples was studied by SEM micro-graphs (SEM, TESCAN), after polishing and etching of the samples, using 5% hydrofluoric acid for 60 sec.

3 . RESULTS AND DISCUSSION

3.1. Crystallization and Foaming Behaviors

Glass composition, heat treatment temperature, soaking time, the amount, and the type of foaming agent are key parameters that affect the final struc-ture of the foamed specimens[1]. It is well known that a soda-lime glass is not very susceptible to crystallization. Surface crystallization of devit-rite(Na2Ca3Si6O16) and wollastonite (CaSiO3) have been reported only in fine particles of it containing a nucleant [10]. On the other hand, the pyroxene groups easily crystallize in the high iron-bearing silicate glasses [16, 18, 22]. Therefore, to enhance crystallization in a soda-lime glass composition, the addition of iron oxide to the glass batches was considered. Being multi-valance, iron oxide can play the role of an oxidizer, and consequently it

Fig.1. X-ray diffraction patterns of the fritted samples.

Table 1. Chemical composition of glasses G20, G24, G28, and G32 (parts weight).

G32G28G24G20

73.8473.8473.8473.84SiO2

31.7027.7424.0319.54Fe2O3

8.778.778.778.77CaO

1.791.791.791.79Al2O3

1.401.401.401.40MgO

12.9512.9512.9512.95Na2O

0.970.970.970.97K2O

0.280.280.280.28TiO2

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

Iranian Journal of Materials Science & Engineering Vol. 17, No. 3, September 2020

3

can provide the oxidizing condition for SiC by the following reactions [3]:

Then, silicon carbide can release CO/CO2 by absorbing the oxygen delivered by the mentioned redox reactions, and/or the oxygen entrapped in the pores as well as the dissolved one in the glass [3, 23].

Figs.1 and 2 show the X-ray diffraction patterns of frit samples and the DTA traces of the glasses G20 and G32, containing the minimum and the max-imum amounts of iron oxide, respectively. As can be seen, the glasses display an amorphous structure in XRD patterns. Also, based on fig.2, the dilato-metric softening point temperature (Td) of G20 and G32 seems to be at approximately 590 and 585°C, respectively. There are two distinct exothermic peaks in each sample which most probably can be attributed to crystallization of glasses. While the first peak does not change with increasing of Fe2O3 and remaining at approximately 735°C, the second one undergoes a significant reduction from 900 to 850°C. Furthermore, the intensity of the latter peak has increased considerably.

Fig. 2. DTA curves of G20 and G32 with a heating rate of 10°C/min.

Fig. 3. XRD patterns of glasses G20 and G32 after heat treatment at 730°C and 900°C for 10 min.

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

4

P. Shahsavari, et al.

To determine the nature of the two peaks, the samples were heat-treated at the two temperatures for 10 min (Fig.3). No crystalline phase was de-tected in the X-ray diffraction patterns of the sam-ples heat-treated at the first exothermic peak for 10 min; but, many spherical micro-crystals were observed in their SEM micrographs (Fig.4-a).

Two pyroxene- based phases including a Fe-bearing diopside (the main crystalline phase) (Fe0.35Al0.2Mg0.44)Ca0.96(Fe0.08Si0.7Al0.2)2O6.12 (JCPDF #08321) as well as acmit and aegirine, were revealed in both glasses after heat treatment at their second DTA crystallization peak. Acmit and aegirine are different polymorphs of NaFe (SiO3)2 (JCPDF #010221) and distinguishing between their XRD patterns was difficult except for some peaks.

These results are consistent with previous re-ports [16, 23-25]. In particular, Salman et al., Kara-manov [24] and Mohamed et al.[16] have shown that high iron-containing silicate glasses mainly crystallize to solid solutions of pyroxenes. It has also been reported that pyroxene phases can be het-erogeneously recrystallized on an intermediate spi-nel-based phase like magnetite. The initial stage of crystallization of iron-rich glasses is characterized by a liquid-liquid phase separation [18]. First, spi-nel-based phases, e.g. magnetite, franklinite, and/or chromite, precipitate within the iron-rich liquid phase. Afterward, they act as the nucleation site for the crystallization of pyroxene during the later heat treatment [16,18]. However, our experiments showed that magnetite was gradually disappeared with increasing of iron oxide.

Based on the XRD results (Fig.3), the intensi-ty of the patterns has significantly increased with iron contents, which is compatible with the DTA results. The dilatometric softening point tempera-ture (Td) of the two glasses is almost the same. However, glass G32 showed a greater tendency to crystallize. This can be attributed to the role of Fe2O3, which a) sometimes plays as a nucleating agent [17, 18, 26], and/ or b) is a component of the crystalline phases that its increasing leads to improvement of crystallization.

Figs. 4-b and 4-c show the cross-sectional views of SiC-free glasses G20 and G32, respective-ly, shaped via the melt casting method, heat-treat-ed at their second peak temperatures for 10 min, and then rapidly cooled. By comparing these two

Glass surface

Fig. 4. SEM micrograph of a) glass G20 after heating at 730°C,mag=10000X; B) glass G20 heat-treated at

900°C,mag=1000X; and C) the glass G32 heat-treated at 850°C for 10 min, mag=1000X.

c

a

b

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

Iranian Journal of Materials Science & Engineering Vol. 17, No. 3, September 2020

5

Figs, it can be concluded that surface crystalliza-tion is the predominant mechanism for crystalliza-tion in both glasses. However, crystallization has been more rapid in G32 than G20, as the dendriti-cally grown crystals are larger in the former one. Based on these results, one can conclude that item (b) is likely responsible for the reduction of DTA crystallization peak temperature with iron oxide.

The addition of SiC (2 wt.%) to G20 and G32, as a foaming agent, led to a decrease in the crystal-lization peak temperature of G32 (Fig.5). It seems that SiC particles has advanced the surface crys-tallization of the mentioned phases via entering new glass surfaces, through foaming of the glass,

and it did not change the types of the precipitated phases(Fig.6).

Fig.7 shows the microstructures of glass-ce-ramic foams G20, G24, G28, G32 after heat treat-ment at 850°C for 25 min. The pores have be-come more isolated and smaller with the increase of Fe2O3. Based on the previous equitation’s, Fe2O3 is known as an oxidizer for SiC particles, and can improve the foaming process by releas-ing of oxygen.

If these reactions had worked in the batch, they would have made the holes bigger. But, it seems that due to the greater crystallization in G32 which usually leads to an increase in glass vis-

Fig. 5 - DTA curves of G20+SiC(2%wt) and G32+SiC(2%wt) with a rate of 10°C/min.

Fig. 6. X-ray diffraction patterns of glasses G20 and G32 containing 2%wt SiC fired at 850°C for 25 min.

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

6

P. Shahsavari, et al.

Fig. 7. SEM micrographs of a) glass G20 and b) glass G24 c) glass G28 d) glass G32 pore morphology after treatment at 850˚C for 25min.

a

c

b

d

a b

Fig. 8. The morphologies of crystalline phases in a) G20; and b)G32 after heat treatment at 850˚C 25min.

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

Iranian Journal of Materials Science & Engineering Vol. 17, No. 3, September 2020

7

cosity [27], its foaming has decreased. It should be noted that the average size of pores in samples G20, G24, G28, and G32 is ~350, 200, 100 µm, and 40 µm, respectively, and the surface pore densie-ty has also decreased with the same trend. Fur-thermore, Instead of slight prismatic crystalline phase which was precipitated in G20(Fig.8a), an extensive precipitation of needle-like small crys-tals are seen in the microstructure of foam pre-pared by G32(Fig. 8b).

3.2. Mechanical Properties

Figs. 8 and 9 depict the variation of porosity and compressive strengths of various SiC-bear-ing specimens as a function of heat treatment temperature and soaking time, respectively.

According to Fig.9, while there is not a mean-ingful difference between the porosity amounts of the specimens fired at 750 °C, the most porosity(90%) and the least one(50%) belong to G20 and G32, respectively, after heat treatment at 950 °C. Furthermore, except for G32, the com-pressive strengths of the other samples are de-creased gradually by increasing the porosity with temperature, as expected, from approximately 70 to about 4 MPa. In G3, increasing the viscosity of the specimen with raising the heat treatment temperature from 850°C to 950°C, due to intensi -fying of crystallization, however, has retarded the reaction in the glass-SiC interfaces. Consequent-ly, the porosity as well as the strength have been remained nearly constant in this heat treatment temperature interval.

Based on Fig.10, the heat treatment time did not affect the porosity, and strength of the samples relative to each other and the G32 acted again differently from G20, G24, and G28 with regard to strength and foaming behavior. Accordingly, it can be conclud-ed that there is a threshold of crystallinity above which the foaming process is prohib-ited. It should also be noted that even the latter glass-ceramic foams show an overall higher strength than the non-crystallized ones investigated in previous studies [2, 8, 9, 11, 13].

Fig. 11(a,b,c, and d) depicts the crushing strengths of various glass-ceramic foams as a function of relative density. It can be seen that the density of specimens G20, G24, and G28 changes almost linearly to higher-value with increasing of their iron oxide. Howev-er, there is not such a trend in G32 such that its relative density limited to 0.4- 0.6 g/cm3

Furthermore, the least and most densities be-long to G20 and G32.

The present authors believe that iron oxide has not helped the foaming process of specimens in this work and it solely plays the role of a crys-talline constituent that improves the mechanical properties of the specimens. This effect originates probably from the high amounts of it used in the compositions.

Yuwei Chi et al.[28] who have investigated the effects of Fe2O3 on the properties of glass

Fig.9. Strength-porosity variation of specimens with heat treatment temperature in the presence of %2wt

SiC for 25 min.

Fig. 10- Strength-porosity variation of specimens with soaking time in the presence of %2wt SiC at 850°C.

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

8

P. Shahsavari, et al.

foams by using iron-containing solid waste have shown that Fe2O3 in glass matrix would increase the viscosity of glass on one hand. On the other hand, they believe an appropri-ate content of Fe2O3 could leading to a better compromise between the glass viscosity and the gas generating, so as to a lower density. Based on their results, adding 1wt. % Fe2O3 as oxidants could promote the foaming ability of the specimen and proper utilization of the Fe3+ ions in parent glass would be beneficial to obtain glass foams with better properties. However, Saeedi Heydari et al. [29] who have

used less than 1.2 wt. % Fe2O3 as an oxidizer in preparation of glass foam from waste cath-ode ray tube display panel (CRT) have reject-ed the role of it as an oxidizer in the glass.

4 . CONCLUSIONS

The present work showed that the addition of Fe2O3 to the SiO2-Na2O-CaO glass system caused the crystallization of pyroxene solid solution crys-tals, i.e. diopside and acimte, within the glasses. Fe2O3 neither acted as an oxidizer for SiC parti-cles nor a nucleant for crystallization. The pyrox-

Fig. 11. Plot of Strength vs relative density for a) glass G20 and b) glass G24 c) glass G28 d) glass G32 foams.

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

Iranian Journal of Materials Science & Engineering Vol. 17, No. 3, September 2020

9

ene phase was precipitated in the glass through the surface crystallization mechanism.

Pyroxene based glass-ceramic foams with high compressive strength (20MPa) and desirable po-rosity (~70%) were prepared using 20 to 28 parts by weight iron oxide, when heat treated at 850°C for 25 min. Using 30 parts by weight iron oxide led to a glass-ceramic with undesirable porosity (45%), but superior compressive strength (60 MPa).

REFERENCES

1. M., Scheffler, and P., Colombo, Cellular ceram-ics, Wiley-Vch, 2006.

2. F., Mear, P., Yot, M., Cambon, and M., Ribes, “Elaboration and characterization of foam glass from cathode ray tubes”, Advances in applied ceramics,2005, 104, 123-130.

3. A. C., Steiner, “Foam glass production from vitrified municipal waste fly ashes”, 2006, 68, 91-118.

4. A. S., Llaudis, M. J. O., Tari, F. J. G., Ten, E., Bernardo, and P., Colombo, “Foaming of flat glass cullet using Si3N4 and MnO2 powders”, Ceramics international, 2009, 35, 1953-1959.

5. J. P., Wu, A. R., Boccaccini, P. D. Lee, M. J. Kershaw and R. D. Rawlings, “Glass-ceramic foams from coal ash and waste glass: produc-tion and characterization”, Advances in applied ceramics, 2006, 105, 32-39.

6. E., Bernardo, and F., Albertini, “Glass foams from dismantled cathode ray tubes”, Ceramics international, 2006, 32, 603-608.

7. P., Colombo, G., Brusatin, E., Bernardo, and G., Scarinci, “Inertization and reuse of waste ma-terials by vitrification and fabrication of glass-based products”, Current Opinion in Solid State and Materials Science,2003, 7, 225-239.

8. E., Bernardo, “Micro-and macro-cellular sintered glass-ceramics from wastes”, Journal of the Euro-pean Ceramic Society, 2007, 27, 2415-2422.

9. H. R., Fernandes, D. U., Tulyaganov, J. M. F., Ferreira, “Preparation and characterization of foams from sheet glass and fly ash using car-bonates as foaming agents”, Ceramics interna-tional,2009, 35, 229-235.

10. D. U., Tulyaganov, H. R., Fernandes, S., Ag-athopoulos, J. M. F., Ferreira, “Preparation and characterization of high compressive strength foams from sheet glass”, Journal of Porous Ma-terials, 2006, 13, 133-139.

11. H., Guo, Y., Gong, S., Gao, “Preparation of high strength foam glass-ceramics from waste

cathode ray tube”, Materials Letters,2010, 64, 997-999.

12. N. M., Bobkova, S. E., Barantseva, E. E., Truso-va, “Production of foam glass with granite sift-ings from the Mikashevichi deposit”, Glass and Ceramics, 2007, 64, 47-50.

13. E., Bernardo, G., Scarinci, P., Bertuzzi, P., Er-cole, L., Ramon, “Recycling of waste glasses into partially crystallized glass foams”, Journal of Porous Materials, 2010, 17, 359-365.

14. E., Bernardo, R., Cedro, M., Florean, S., Hreg-lich, “Reutilization and stabilization of wastes by the production of glass foams”, Ceramics international, 2007, 33, 963-968.

15. S., Hasheminia, A., Nemati, B. E., Yekta, P., Al-izadeh, “Preparation and characterization of di-opside-based glass-ceramic foams”, Ceramics International, 2011, 38, 2005-2010.

16. E., Mohamed, P., Shahsavari, B., Eftekhari-Yek-ta and V., K.Marghussian,” Preparation and Characterization of Glass Ceramic Foams Pro-duced from Copper Slag, Trans. Ind. Ceram. Soc., 2015, 74, 1-5.

17. R. K., Singh, G. P., Kothiyal, A., Srini-vasan, “Influence of iron ions on the magnet-ic properties of CaO-SiO2-P2O5-Na2O-Fe2O3 glass-ceramics”, Solid State Communications, 2008,146, 25-29.

18. S. M., Salman, S. N., Salama, “Pyroxene solid solutions crystallized from CaO-MgO (Li2O, Fe2O3)-SiO2 glasses”, Ceramics international, 1986, 12, 221-228.

19. P., Alizadeh, B., Eftekhary Yekta, A., Gervei, ”Effect of Fe2O3 addition on the sinterability and machinability of glass-ceramics in the sys-tem MgO–CaO–SiO2–P2O5”,Journal of the Eu-ropean Ceramic Society, 2004, 24, 3529-3533.

20. W., Holand and G. H., Beall, “Glass-ceramic technology, Wiley-American Ceramic Society, Ohio, 2012,118.

21. V. G., Sister, E. M., Ivannikova, M. A., Se-min, and A. A., Egorov, “Preparation of highly porous cellular glass materials in the field of pyroxene crystallization”, Chemical and Petro-leum Engineering, 2011, 46, 631-633.

22. P. A., Bingham, J. M., Parker, T., Searle, J. M., Williams, K., Fyles, “Redox and clustering of iron in silicate glasses, Journal of Non-Crystal-line Solids, 1999, 253, 203-209.

23. S. N., Salama, S. M., Salman, “Crystalliza-tion characteristics of iron-containing spodu-mene-diopside glasses”, Journal of the Europe-an Ceramic Society, 1993, 12, 61-69.

24. A., Karamanov, M., Pelino, “Crystallization

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]

10

phenomena in iron-rich glasses”, Journal of non-crystalline solids, 2001, 281, 139-151.

25. Z. j., Wang, W., Ni, K. Q., Li, X. Y., Huang, L. P., Zhu, “Crystallization characteristics of iron-rich glass-ceramics prepared from nickel slag and blast furnace slag”, International Journal of Minerals, Metallurgy, and Materials, 2011, 18, 455-459.

26. I., Levitskii, “The Effect of Iron Oxides on the Properties and Structure of Glazed Glasses”, Glass and Ceramics, 2003, 60, 111-114.

27. B., Eftekhari Yekta, V. K., Marghussian, “Sin-tering of β. q. SS and gahnite glass ceramics”, Journal of the European Ceramic Society, 1999, 19, 2963-2968.

28. Yuwei Chi Lin, J. and Xu, B., “Effects of Fe2O3 on the Properties of Glass Foams Prepared by Iron-Containing Solid Waste”. Glass Phys. Chem., 2019, 45, 104-110.

29. Saeedi Heydari, M., Mirkazemi, S. M., and Abbasi, S., ”Influence of Co3O4, Fe2O3 and SiC on microstructure and properties of glass foam from waste cathode ray tube display panel (CRT)”, Journal Advances in Applied Ceramics Structural, Functional and Bioceramics, 2014, 113, 173-179.

P. Shahsavari, et al.

Dow

nloa

ded

from

ijm

se.iu

st.a

c.ir

at 3

:53

IRD

T o

n S

atur

day

July

24t

h 20

21

[ D

OI:

10.2

2068

/ijm

se.1

7.3.

1 ]